BIRD ALARM CALLS

Some examples will serve as illustrations. Many small birds have two types of alarm call, one the "hawk-alarm call," the other a mobbing call; the former is often difficult to locate, the latter easy (Marler, 1955). The ventriloquial type is given especially on seeing a flying hawk. The locatable type is used with either a ground predator (such as a fox) or a stationary aerial predator (such as an owl perched in a tree). One might think of these two classes of calls as names for two classes of predators, but other factors complicate the issue. A male chaffinch, the species in which I first discovered this dichotomy between ventriloquial and locatable calls, will also give the "hawk-alarm call" in other situations. At the height of the breeding season, for example, he may give it if a person comes close to the nest containing his young.

At this point a sceptic, convinced of the inflexible automaticity of the animal as an unthinking machine, is likely to respond that this is hardly surprising if, as is surely appropriate, one views these alarm calls not as analogs for "symbolic" names but rather as "affective" signals, such as we generate in unconscious signs of emotion in voice, gesture, and expression. If there are two calls, so the thinking goes, they probably represent different levels of arousal. As I interpret this line of argument, two facts remove these calls from the naming process as manifest in our language. One is the lack of specificity in circumstances surrounding production of the "hawk-alarm call." The other is the apparent correiation with arousal levels, intense with the "hawk-alarm call," less so with the mobbing call.

The focus in this interpretation is on the denotative aspect of meaning. If we were to allow ourselves to entertain the notion that calls can have connotational meanings for animals (Griffin, 1976 ), then the lack of specificity could be viewed as in harmony with a more general meaning for the call than one particular class of predator. It might, for example, represent the abstract idea of a certain class of dangers, a class meriting a particular mode of response in those companions the signaler is presumably helping by its communications. The response of suddenly freezing in a crouched, cryptic position is as appropriate for nestling chaffinches threatened by a mammalian predator as it is for foraging adults threatened by a hunting hawk.

ALARM CALLS OF MONKEYS

Struhsaker (1967) discovered a remarkable array of alarm calls in the vervet monkey (Table 1). Inspection of the situations in which the four adult male alarm calls are uttered shows that they can be arranged in a series of increasing specificity. The first two, the "uh" and the "nyow," are evoked by rather generalized stimuli, and seem to function primarily to alert others. The third, the "chutter," is associated with either a man or a snake, especially the latter. The fourth, the "threat-alarm-bark," the only one that is the sole prerogative of the adult male, is given on sighting a major predator. It evokes not approach and mobbing, as is the case with a snake, but precipitant flight to cover. This series can also be interpreted in terms of an increasing level of arousal.

Table 1. Alarm calls of the vervet monkey, stimuli evoking them, and responses of others. After Struhsaker (1967) and personal communication.

Such an interpretation is less straightforward with the alarm calls of the female vervet monkey, who uses all but one of the series. The "rraup" and the "chirp," for example, are given in rather specific circumstances, and evoke specific and contrasting responses—descent from the tree-tops in one case and ascent into the treetops in the other. It is not clear which is associated with a higher level of arousal. Perhaps here, more than anywhere else in vervet vocal behavior, one might begin to think of these two female alarm calls as serving as names for particular referents. As with the chaffinch, however, each alarm call has more than one "denotational" referent.

In spite of this multiplicity of referents, may there not be "connotational" meanings to the vervets themselves with more unity than is immediately evident to us? One might think of alarm calls, not so much as having a large, ill-defined class of denotations, but rather as representing the idea of a certain class of dangers favoring a particular escape strategy. In our own language, after all, the lack of one specific class of referents by no means disqualifies a signal from a naming function.

Even while entertaining this interpretation, there remains the correlation of different vervet alarm calls with independent signs of varying levels of arousal. Such correlations, regular and orderly, are well documented in other kinds of primate signaling behavior (e.g., Green, 1975; Eisenberg, 1976). Still it is not clear to me whether the arousal associated with much animal signaling is best viewed as conflicting with a symbolic function or as supplementing it.

CALLS OF THE CHIMPANZEE AND THE GORILLA

Some primate vocalizations are notable for the relatively low arousal level with which they are associated. Chimpanzees have a vocalization I call "rough grunting," associated specifically with the discovery and eating of a highly preferred food such as a bunch of ripe palm nuts or, in an experimental situation, a bunch of bananas (Marler, 1976). Gorillas have a call with a similar acoustical morphology, labeled "belching" by Fossey (1972 ). While gorilla "belching" does occur during eating, it also occurs in a wide variety of other situations associated by Fossey with quiet relaxation, either while resting or while moving slowly through forage and feeding. The contagious chorusing of "belching," with no counterpart in the chimpanzee, apparently aids the gorilla group in maintaining contact. The referential designata for gorilla "belching" are more generalized than those for the "rough grunting" of the chimpanzee.

When we compare the social organization of these two species, a correlation with usage of the two calls emerges. The close-knit and coherent social unit of the gorilla is associated with the use of belching for maintaining contact while moving and resting in dense cover. The chimpanzee, with its much looser pattern of social organization, uses its version of the call in a more specific situation, thus assuming more of the function that we associate with a name. It is not hard to imagine the evolutionary transition from one condition to the other as social demands upon the vocalization vary with shifts in the pattern of social organization. Variations in referential specificity may thus be viewed as adaptive. This is true in our own languages, where a class of objects of indifferent detail for one culture, and a single name, may be subdivided into a multiplicity of separately named subclasses by a culture in which the added specificity assumes ecological importance.

AFFECTIVE STATES AND AROUSAL

It is not clear from the evidence on chimpanzee and gorilla behavior whether the two calls are associated with different levels of arousal. In both the levels seem to be low. But even if they were strikingly different, one can view such variations not as confounding the function of naming but rather as supplementing it. It potentially adds a valuable new communicative dimension. I can best illustrate what I have in mind with a quotation from Norbert Wiener (1948 ).

With no other signaling elements than signs of arousal level and the deictic property of eyes and where they are looking, such behavior has rich communicative potential. The theme that we have underestimated the rich communicative potential of affective signaling is also discussed by Premack (1975 ). He opens a chapter on the origins of language as follows :

He goes on to develop the difference between the two kinds of signaling, referential (= symbolic) and affective (= excited or aroused), suggesting that information of the first kind consists of explicit properties of the world next door; information of the second kind, of affective states, which he assumes to be positive or negative, and varying in degree. He goes on:

Premack makes the point that as long as there is some concordance between the preferences and aversions of communicants, a remarkable amount of information can be transmitted by an affective system. While he explicitly restricts his discussion to "what" rather than "where," one may note, harking back to the Wiener quotation, that incorporation of a deictic component in the signal—pointing or looking—not only indicates where, but also adds a highly specific connotation to the information—not just any apple tree, or all apple trees in the abstract, but this particular one.

FOUR DIMENSIONS OF "AROUSAL"

One drawback with arousal models of motivation is that although compelling in some respects (Mason, 1967, 1971) they leave unexplained a lot of detail of the natural behavior of animals. What are the simplest steps that could be taken to increase the efficiency of affective signaling, to specify its content more precisely? In Wieners illustration there is one dimension. In Premack's there are already two. I tentatively suggest adding at least two more (Table 2).

Table 2. Four proposed dimensions of "affective" signaling

First, there is a level of arousal. Second, there is some indication of positive or negative sign, showing whether the signaler is inclined to approach the stimulus object or "referent" or to withdraw from it and opening up a similar set of alternatives for a responding companion. Beyond this point, already reached by Wiener and Premack, it would be advantageous if the signaler could give even a general indication of the class of referent being signaled about. Otherwise a hungry animal, for example, might approach a referent indicated by a companion as worthy of positive affect only to discover that it indicated not food but a social companion or a mate. Much time and effort would be saved if, at the very least, the signaler could give some indication as to whether the referent is environmental or social in nature. Environmental designata might include hazards such as predators or bad weather or resources such as food and drink or a safe resting place. Social referents, again with either positive or negative connotations, might include a mate or companion, an infant in need of care, all worthy of approach, or perhaps an enraged male of the species, who is to be avoided.

Intriguingly, some of the data on human emotional states and their classification are not incompatible with the four dimensions I have indicated. Fig. 1 is modified from Plutchik (1970), a psychiatrist interested in measuring the behavioral biases of different kinds of disturbed patients. His major emotional categories, which can be viewed in part as communicative devices, are shown, with minor changes, in the center of the diagram. In the outer ring I have added some equivalent ethological categories of ongoing behavior. Such a classification, crude though it is, encompasses most of the major sets of activities that constitute the ethogram of an animal. It is also compatible with the four affective dimensions of arousal/depression (A), approach/withdrawal (B), social engagement/disengagement (C), and object acceptance/rejection (D).

Fig. 1. Human emotional states (inner ring) with some equivalent ethological activities (outer ring) arranged according to whether their connotations are positive, negative, or neutral.

A-A = a possible arousal/depression dimension

B-B = a locomotor approach/withdrawal dimension

C-C = a social engagement/disengagement dimension

D-D = an object acceptance/rejection dimension. Modified after Plutchik (1970).

The details of these speculations are subordinate to the general point I am making—that, in some communicative circumstance there would be benefits to an increasing degree of specificity in the multiple relationships between production of a signal, its underlying physiological basis, and the kind of referential events contingent to its production.

"SPECIFICITY" IN SYMBOLIC AND AFFECTIVE SIGNALING

I believe that this theme of specificity is central in differentiating between what we are accustomed to think of as symbolic signaling on the one hand, and "affective" signaling on the other. As I have indicated in Table 3 there are at least four related aspects to this specificity.

Table 3. Features of "specificity" differentiating "symbolic" from "affective" signaling

1. Referential specificity (to the class of objects, events or properties represented)

2. Motor specificity (relation to other responses to the referent)

3. Physiological specificity (association with arousal and emotion)

4. Temporal specificity (coupling to immediate referential stimulation)

Symbols have a certain referential specificity. Thus to the extent that the vervet monkey chutter may be given to a man or a snake some no doubt feel that it fails to satisfy one of the conditions for a symbolic signal. However it may be a mistake to assume that vervet monkeys classify referential events or think about them, in our terms. Comparative studies of human cognition reveal many cultural disparities in the ways in which we view the world.

Another aspect of symbolic signaling is motor specificity. We generally require at least a potential ability to perform the motor act of signal production separately from the other behavioral responses that are associated with stimuli from the particular referent in question—an alarm call dissociated from acts of fleeing, a food call separate from eating. A third issue is physiological specificity, the potential dissociation from whatever degrees and qualities of arousal and emotion are associated with response to the referent. And these two relate in turn to what might be called temporal specificity, the fact that affective signaling, at least in animals, is usually closely bound in time to immediate or recent perception of the referent.

Signalers must themselves survive and reproduce, and a dissociation of signaling from other aspects of physiological and behavioral responding that contribute to individual fitness is likely to be dysgenic in most circumstances. Without temporal displacement of referential experience from signaling about that experience I find it hard to imagine the dissociation being other than disadvantageous.

There is one class of behavior in which such dissociation occurs regularly in animals, even constituting a useful identifying criterion. I am referring to play behavior, which often involves reorganization of the normal temporal relationships of signals, their usual referents, and the other behaviors typically accompanying them (Loizos, 1966). Thus such dissociation is not a physiological impossibility for animals. It may best be thought of as something that in most circumstances animals can well do without, other than in play, where youthful experimentation free of affective accompaniments may have beneficial ontogenetic consequences.

The point I am striving to establish is that symbolic and affective signaling should be thought of as differing in degree rather than kind. The particular constellation of relationships between referential, motor, and physiological specificity manifest in much animal signaling, often affective in nature, should not be thought of as something that animals do because they are physiologically incapable of going further in the direction of symbolic signaling. Rather it represents an adaptive compromise that, even in animals, can be seen to move to and fro along a continuum of specificity depending on which particular solution solves a given behavioral problem most efficiently. Furthermore, though animals are highly biased toward the "affective" extreme, they probably exhibit some simultaneous intermingling of "affective" and "symbolic" signal elements, much as intonation and gesture supplement the more symbolic components of our own communicative behavior.

REFERENCES

Cherry, C. 1966. On Human Communication. Cambridge: MIT Press.

Eisenberg, J. F. 1976. "Communication Mechanisms and Social Integration in the Black Spider Monkey Ateies fusciceps robustus, and Related Species." Smithsonian Contributions to Zoology 213:1-108.

Fossey, D. 1972. "Vocalizations of the Mountain Gorilla (Gorilla gorilla beringet)." Anim. Behav. 20:36-53.

Green, S. 1975. "Variation of Vocal Pattern with Social Situation in the Japanese Monkey (Macaca fuscata): A Field Study." In Primate Behavior, vol. 4, L. A. Rosenblum, ed. New York: Academic Press.

Griffin, D. R. 1976.The Question of Animal Awareness. New York: Rockefeller University Press.

Loizos, C. 1966. "Play in Mammals.^,י Symp. Zool. Soc. London vol. 18. In Play, Exploration and Territory in Mammals, P. A. Jewell and C. Loizos, eds. New York: Academic Press.

Marier, P. 1955. "Characteristics of Some Animal Calls.״Nature (London) 176:6-8.

______. 1976. "Social Organization, Communication and Graded Signals: Vocal Behavior of the Chimpanzee and the Gorilla." In Growing Points in Ethology, P. Bateson and R. A. Hinde, eds. London: Cambridge University Press.

Mason, W. A. 1967. "Motivational Aspects of Social Responsiveness in Young Chimpanzees." In Early Behavior: Comparative and Developmental Approaches, H. W. Stevenson, E. H. Hess, and H. L. Rheingold, eds. New York: John Wiley and Sons.

______. 1971. "Motivational Factors in Psychosocial Development." In Nebraska Symposium on Motivation, W. J. Arnold and M. M. Page, eds. Lincoln: University of Nebraska Press.

Palmer, F. R. 1976.Semantics. London: Cambridge University Press.

Plutchik, R. 1970. "Emotions, Evolution, and Adaptive Processes." In Feelings and Emotions, M. B. Arnold, ed. New York: Academic Press.

Premack, D. 1975. "On the Origins of Language." In Handbook of Psychobiology, M. S. Gazzaniga and C. B. Blakemore, eds. New York: Academic Press.

Quine, W. V. О. 1953.From a Logical Point of View. Cambridge: Harvard University Press.

Sebeok, T. A. 1968.Animal Communication. Bloomington: Indiana University Press.

______. 1977.How Animals Communicate. Bloomington: Indiana University Press.

Struhsaker, T. 1967. "Auditory Communication among Vervet Monkeys (Cercopithecus aethiops)." In Social Communication among Primates, S. A. Altmann, ed. Chicago: University of Chicago Press.

Wiener, N. 1948.Cybernetics. New York: John Wiley and Sons.